Troubleshooting tools

In this section, we will mention several tools and utilities that can help the application developer identify system bottlenecks and performance issues. (Note that here, to save space and time, we do not delve into the dozens of usual suspects—well-known system monitoring utilities on Linux such as ps, pstree, top, htop, pidstat, vmstat, dstat, sar, nagios, iotop, iostat, ionice, lsof, nmon, iftop, ethtool, netstat, tcpdump, wireshark—and instead mention the newer ones). Here is an important thing to remember when performing data collection (or benchmarking) for later analysis: take the trouble to set up a test rig and, when using it, change (as far as is possible) only one variable at a time for a given run so that you can see its impact.

perf

...